1. Field of the Invention
The present invention relates to an image processing apparatus and an image processing method, and in particular, relates to a technique that reduces noise while preserving an edge of an image.
2. Description of the Related Art
In the field of imaging devices such as a camera, there is a case where digital image data acquired through an imaging element (such as a CCD or a CMOS) includes a noise component, and various methods of removing this noise component are proposed.
For example, Japanese patent application publication No. 2007-288439 discloses an image processing apparatus that applies noise reduction processing to a color image of RGB. According to this image processing apparatus, the distance between a target pixel and a peripheral pixel therearound on the RGB color space is calculated, a weighting coefficient for each peripheral pixel is calculated according to the distance on the color space, and image data of peripheral pixels in addition to the target pixel is subjected to weighted average processing to calculate the pixel value of the target pixel.
Moreover, Japanese patent application publication No. 2002-222416 discloses an image quality improvement apparatus that removes noise from an image. According to this image improvement apparatus, covariance data between color channels in image data is calculated, a filter which minimizes the mean square error in the k-dimensional color space with an original image that does not deteriorate by noise is derived according to this covariance data and covariance data between color channels of prescribed noise, and noise reduction processing is performed by this derived filter.
Moreover, Japanese patent application publication No. 10-112803 discloses an image processing apparatus that removes noises by use of a reference value recorded in a reference value buffer. According to this image processing apparatus, an output pixel value is determined by weight calculation with respect to an input pixel value.
Various methods related to noise reduction/removal are proposed as described above, but each method has difference in noise reduction effects, and in such a case where the noise level is relatively large (high), an excellent noise reduction effect cannot be necessarily acquired.
For example, Japanese patent application publication No. 2007-288439 discloses an image processing method according to which the distance on the color space is calculated after low-pass filter processing is performed. In a case where only the low-pass filter is simply applied, the edge preservability may become poor though it is still effective for the noise reduction effect. In particular, in a case where the noise level is large, defects such as color spread (blur) and color loss in a color boundary part of an image may occur.
Moreover, even in the image improvement apparatus of Japanese patent application publication No. 2002-222416, when noise of a large level is included in image data, the separation performance between the image signal component and the noise component deteriorates. As a result, in a case where the image improvement apparatus of Japanese patent application publication No. 2002-222416 is used, strong noise may remain and color spread may occur in a color boundary part.
Moreover, in the image processing apparatus of Japanese patent application publication No. 10-112803, since weight averaging is performed on the basis of only an input pixel value and a reference value, a resultant noise reduction effect is inherently small.
Therefore, it is desired to propose a new technique of effectively removing image noise while preventing color spread or the like in a color boundary part and preserving the original edge (boundary part) sharpness of an image even in a case where a noise component of a large level is included in image data.